Proceedings -- Parallel Computing.

The research reports presented in this volume focus on the implications of the T9000 microprocessor, which offers new elements in transputing and parallel programming. Subjects discussed include genetic algorithms, image analysis, neural networks, robotics and parallel architectures.

During the past fifteen years concurrency in programming languages such as Java rose and fell, and again became popular. At this moment developers advise us to avoid concurrency in programming. They are using a host of deprecated methods in the latest releases How are we to understand the love-hate relationship with what should be a widely used approach of tackling real-world problems? The aim of rchitectures, Languages and Techniques is to encourage the safe, efficient and effective use of parallel computing. It is generally agreed that concurrency is found in most real applications and that it should be natural to use concurrency in programming. However, there has grown up a myth that concurrency is "hard" and only for the hardened expert. The papers collected in this book cover the whole spectrum of concurrency, from theoretical underpinnings to applications. The message passing style of concurrency, developed in the Communicating Sequential Processes (CSP) approach, is considered, and extensions are proposed. CSP's realization in the programming language occam is used directly for applications as diverse as modeling of concurrent systems and the description of concurrent hardware. This latter application may be compared to the use of Java for the same purpose. Concurrency and the use of Java is the subject of further papers, as is the provision of CSP-like facilities in Java and C and techniques to use these languages to construct reliable concurrent systems. At a time when concurrency gives headaches, this book brings a welcome breath of fresh air. Concurrency can really be a positive way forward.

ZEUS (Centres of European Supercomputing) is a network for information exchange and co-operation between European Supercomputer Centres. During the fall of 1994 the idea was put forward to start an annual workshop to stimulate the exchange of ideas and experience in parallel programming and computing between researchers and users from industry and academia. The first workshop in this series, the ZEUS '95 Workshop on Parallel Programming and Computation, is organized at Linkoping University, where the Swedish ZEUS centre, NSC (National Supercomputer Centre) is located. This is open for all researchers and users in the field of parallel computing.

Abstract Machine Models have played a profound though frequently unacknowledged role in the development of modern computing systems. They provide a precise definition of vital concepts, allow system complexity to be managed by providing appropriate views of the activity under consideration, enable reasoning about the correctness and quantitative performance of proposed problem solutions, and encourage communication through a common medium of expression. Abstract Models in Parallel and Distributed computing have a particularly important role in the development of contemporary systems, encapsulating and controlling an inherently high degree of complexity. The Parallel and Distributed computing communities have traditionally considered themselves to be separate. However, there is a significant contemporary interest in both of these communities in a common hardware model; a set of workstation-class machines connected by a high-performance network. The traditional Parallel/Distributed distinction therefore appears under threat.

This volume contains papers presented at the 18th meeting of the World Occam and Transputer User Group (Wotug). The papers cover a wide range of transputer and OCCAM-related topics, such as the the porting and development of the OCCAM language (highlighting the need for cross platform implementations of OCCAM compilers), design approaches and applications.

Technical articles on the 19th meeting of the World Occam and Transputer User Group (WoTUG). They cover a wide range of topics from hardware applications to software tools for of parallel processing support; not solely related to transputers. Part of the book focusses on the retargeting of the occam compiler to a range of other processors.

This volume includes very high quality papers in different areas of computer and information sciences. The main themes are (computer network) performance evaluation and artificial neural networks and their applications. The latest developments in these areas are presented by a number of distinguished researchers from all over the world. These proceedings of The 13th International Symposium on Computer and Information Sciences (ISCIS'98) contain outstanding papers specifically related to the areas of "Gelenbe" neural networks and their applications, performance of computer-communication networks, simulations and analytic methods in order to study the performance of telecommunication networks, scheduling and resource allocation in computer and multimedia systems, stochastic ordering applied to performance evaluation, and simulation of virtual humans.

The 21st century will be the age of network computing. Among the many key technologies in this field, parallel computing and networking technology will play very important roles. In this book emphasis is placed on networking and modeling parallel computing. The topics cover parallel computing algorithms, parallel software, massively parallel computing systems and related applications. Articles cover parallel computing, networking and related applications, to initiate discussions. Since the appearance of Transputer chip T9000, C104, and standardizations of IEEE1355, Transputer systems seem to have opened a new interesting area of parallel computing, networking and many practical applications.

The Network Designer's Handbook will help anyone trapped between limited bandwidth, fault-intolerant computer buses and expensive, over-engineered telecommunications technology. It will help anyone looking for new cost-effective ways to build LAN switches, RAID systems, multimedia servers or multiprocessors. It will help the small company looking for an edge to break into the market, and the large one looking for ways to improve its margins and boost its market share. This handbook, the result of over six man-years of effort at the PACT Research Institute, provides solid engineering data for the computer systems professional. Four cpu-months of system simulation are summarised in an easy-to-read form, allowing the consequences of different design decisions to be simply compared. The Network Designer's Handbook explains the principles of the new generation of small-scale low-buffer serial interconnects. Using the specific example of IEEE 1355-1995 links and the STC104 high-valency switch chip it shows how this technology can provide modular, fault-tolerant and scalable interconnect. The picture is rounded out with descriptions of network topologies, case studies and many practical tips.